Crosscloud is a decentralized system, still being developed, which is intended to give users the freedom to move comfortably between competing cloud-based applications. Where today users are locked in by sites which hold onto their data and their social connections, when crosscloud (or something like it) is widely deployed, users will be free to switch to new applications in search of better features, pricing, and policies.
The flagship Crosscloud application is cimba, a microblogging tool being built at MIT CSAIL. Cimba users get an experience like that provided by Twitter, Facebook news feeds, etc, but they can switch to another Crosscloud microblogging app and still see the same social network with the same posts and the same metadata. More impressively, they can also switch vendors for their back-end data storage with no impact on user experience for themselves or their contacts. (At least that's the vision; it's not there yet.) Many other Crosscloud applications are possible, spanning the whole range of Internet success stories and dreams of a connected world (see potential application areas).
Freedom of choice in the market may seems like a goal in itself, but its real value becomes clear when we consider the benefits that will come with that freedom:
Innovation. When your data and social connections are locked in to one platform or application, it's hard to try out interesting alternatives. Even if you try them, it's hard to make the switch, because all your history and previous work is in your old system. Your social connections are on the old system, too. Can you get all your friends or colleagues to move at the same time? And if people like you don't switch, new systems will fail for lack of users. When users are locked in, most serious innovation is locked out.
Connection. For a variety of reasons, certain people you know probably can't or won't use any particular social networking platform. This makes it hard to connect with them online, whether it's for work or just for fun. With a large enough group of people (sometimes only 5 or 10) it's impossible to find a platform they all use. In contrast, With crosscloud software, people will have a wide variety of vendors and platforms to choose from. As long as each person can find one that works for them, you can all be connected.
Stability. You can keep using the same version of the software as long as someone is willing to support it. No more being forced to accept every product change or give up the product.
Security. Because crosscloud products never lock you are in, you are free to seek out providers who offer high levels of data security. Because everyone is free to do this, we expect in time the market will provide many strong options.
Privacy. Although some invasions of privacy are due to failures in data security, many are part of routine business on the Internet today. Advertising supported products, in general, coerce users to grant intimate access to their data. Crosscloud systems don't prevent such access, but they do give users freedom to choose alternatives.
Reliability. A system providing crosscloud functionality has to be engineered without introducing any single point of failure (since control over that point would mean control over the system, and scaring away users and ecosystem participants). So the overall system is likely to be as reliable as the Internet in general. Individual servers might be unreliable, but since users can easily switch, there should be strong market pressure toward reliability.
Customizability. Also known as "hackability" or "modability", the ability of a system to be tinkered with and modified to suit particular needs is important to adoption and user satisfaction. Centralized systems have a much harder time supporting this.
Cost Savings. Because the barriers to entry for competitors are much lower, prices are likely to be lower.
To make the crosscloud vision a reality, three main things have to happen:
There are some open questions in computer science about whether it's feasible to build a decentralized system of this scale, with these characteristics.
So we need to build demonstration apps (eg cimba) and publishing any new science we do in the process. This is: prove that it can be built by building it.
Building a system with the right properties is not enough. It has to be something that is easy enough to implement that it ends up widely deployed, as the only system like this. Imagine if instead of the Internet, we had several different global networks, and you could only reach the people on the same one. (Some of us remember those days, with BITNET, uucp, etc.)
This means people working toward the crosscloud vision need to work together to reach consensus on the protocols. And during the time before consensus is reached, people will need to keep updating their software. At some point along the way, Crosscloud may need to merge with similar systems for the benefit of the users.
At this point, the Linked Data technology stack is being used and extended to build crosscloud systems, with standards work being done at W3C. We view the crosscloud vision as one of the driving use cases for Linked Data.
Even if the technology works well and there is complete consensus on the standards, if there aren't enough applications or service providers, users wont have the intended level of freedom and benefits.
So it's important to make sure there are plenty of people able and willing to build crosscloud applications and run crosscloud servers.
Crosscloud development is led by Sandro Hawke and Tim Berners-Lee at MIT and involves various projects around the world. Related projects that are not part of the Crosscloud effort, but might converge eventually, are being listed on the reading list.